What we’ve done so far

We wish we could put everything we’ve done over the years of engineering on this page. One day – when we get a little less busy – we’ll write a memoir. Hopefully.

Recent projects

Here is where we store proofs of our experience developing challenging projects.


An order processing app for restaurants

  • 1.6M 1.6M seed investments
  • 500 000 orders processed


A personal running coach for improving cardiovascular system

Real-time high-precision location tracking
BTLE device connection
Kalman Filter


App for famous art gallery
PDF reader
EPUB reader
Large files download


The first social network dedicated to shopping. Share your style with your friends, look for products you love and shop directly within the app.
  • €350k seed round funded by Lucey Fund
  • 1st Round investment in 2018

Kabuni App

Kabuni‘s platform enables home décor designers worldwide to collaborate with clients
  • AU$2.2 million seed investment
  • Total $12 million investment in 2 rounds


A crowdsourced navigation
and traffic app for truckers
  • 1500 active truck drivers on start


Slap is the most effective way to keep your fitness and weight loss resolutions
HealthKit integration
Fitbit integration
MapMyRun integration
PayPal integration
Stripe integration
Rich animation

Secure My Health

Manage medical prescriptions and remind patients to purchase medications on time
Medicine database
Color blindness friendly
Rich iOS accessibility support


A mobile wallet app
for a global settlement system
  • 2M active users a month

Our wins

Some engineering tasks require a separate block. Like this one.

Challenge: Text-based emotion detection

Artificial intelligence • Tensor Flow • Apple Core ML • Support Vector Machines algorithm • Naive Bayes algorithm • Stemming

Emotion detection can be used by online businesses to improve the quality of their customer support, to understand their consumers better, and to make their marketing messaging more effective.

To detect emotions in the text our algorithm uses several levels of classifiers and filters. Here is how it works: the text is first processed with a syntactic parser. From here, we need to filter all the context surrounding a word or a phrase to understand its meaning. That's semantic analysis. Then the system classifies conversations into happy, sad, angry, resentful, and other types of feelings.

Our algorithm is incredibly shrewd when it comes to sarcasm. Humans aren't very good at separating sarcasm from sincerity online. They can guess that a sentence is sarcastic with 62% of accuracy. Our algorithm reaches an overall accuracy of 71%.

See more details

Challenge: Nudity detector

Artificial intelligence • Tensor Flow • Apple Core ML • OpenCV image correction • Object recognition

Nudity detector is an algorithm that automates content moderation for images. It can detect unacceptable illustrations, and can recognise nudity or other sexual content on 3D computer graphics.

In order to keep their apps clean, online companies employ entire departments to moderate content posted by their users. Our algorithm works faster than humans and doesn't demand monthly paychecks.

However, just because the color of skin prevails in the picture, it doesn't mean that the image posted by the user contains inappropriate content. This might just be a naked...hand. Our algorithm uses additional filters that recognize the differences in skin tone and can identify an object in the picture.

Certain nude pictures that are acceptable in one country or culture may not be very suitable in other countries or cultures. Our nudity detector understands regional and cultural nudity norms. When moderating content the algorithm takes into account the target audience's culture, age, and gender.

The more images the algorithm processes, the better it gets.


Challenge: Real time speech emotion and sentiment recognition

Artificial intelligence • Tensor Flow • Apple Core ML • Support Vector Machines algorithm • Naive Bayes algorithm • Stemming • Google VTT • Logistic Regressions (Legit) • Hidden Markov Networks

This algorithm is an extended version of the text-based emotion detection system.

The algorithm uses voice tone analyzer to recognize emotions in speech. In other words, it can understand not only what was said, but also how it was said.

Using this tool, companies can evaluate the effectiveness of their call centers. They can analyse customer conversations and use the results of this analysis to improve sales and customer service.


Let's do it

Tell us you want to work together
and we'll take it from there
Get in touch